Granted Patent B2

US 12,335,502 · App. 18/136,091 · Granted Jun 17, 2025

Inventors: Lien-Fei Chen (Hsinchu, TW); Guichun Li (San Jose, CA); Xiang Li (Saratoga, CA); Shan Liu (San Jose, CA)

Assignee: TENCENT AMERICA LLC

H04N19/44H04N19/105H04N19/132H04N19/176H04N19/70

Abstract

Claims (84)

1. A method for video decoding, comprising:

receiving a video bitstream comprising a current block in a current picture;

decoding a value of a syntax element associated with the current block in the current picture, the syntax element indicating whether the current block is coded with a geometric partition mode (GPM) along a partition edge intersecting the current block; and

when the value of the syntax element indicating that the current block is coded with the GPM and a condition for adaptive blending is satisfied,

reordering width candidates in a width candidate list using template matching (TM) that is based on a current template of the current block and reference templates that correspond to the respective width candidates;

selecting a width candidate from the reordered width candidate list;

determining a width of a blending area based on the selected width candidate, the blending area surrounding the partition edge as defined by boundaries on both sides of the partition edge, the boundaries being parallel to the partition edge, the width of the blending area measured perpendicular to the partition edge;

determining the blending area based on the width of the blending area; and

reconstructing samples within the blending area in the current block by applying the adaptive blending using the determined blending area.

2. The method of claim 1 , wherein the reordering the width candidates comprises:

extending the partition edge from the current block into the current template, an extended portion of the partition edge being in the current template;

for each width candidate in the width candidate list,

determining a reference template corresponding to the respective width candidate based on the GPM and the current template, reference samples within a template blending area that surround the extended portion of the partition edge being determined based on a blending process, the template blending area being in the reference template, a width of the template blending area being based on the respective width candidate; and

determining a TM cost corresponding to the respective width candidate based on the current template and the reference template; and

reordering the width candidate list based on the determined TM costs.

3. The method of claim 1 , further comprising

when the condition for the adaptive blending is not satisfied,

determining the width of the blending area based on the width candidate list without reordering the width candidates in the width candidate list; and

reconstructing the samples within the blending area in the current block by applying a blending process.

4. The method of claim 3 , wherein

coding information of the current block indicates whether the condition for the adaptive blending is satisfied.

5. The method of claim 3 , wherein

when an extended portion of the partition edge intersects the current template,

an area between a first boundary and a second boundary that are parallel to the partition edge includes a first area and a second area, the area including the extended portion of the partition edge, the first boundary and the second boundary being on opposite sides of the partition edge and having a distance of the largest width candidate in the width candidate list from the partition edge,

the first area is an area between the first boundary and the partition edge that overlaps with the current template,

the second area is an area between the second boundary and the partition edge that overlaps with the current template,

the condition for the adaptive blending is satisfied based on the first area being equal to the second area; and

the condition for the adaptive blending is not satisfied based on the first area being not equal to the second area; and

when the extended portion of the partition edge does not intersect the current template, the condition for the adaptive blending is not satisfied.

6. The method of claim 3 , wherein

the condition for the adaptive blending is not satisfied when an above template above the current block or a left template to the left of the current block is not available.

7. The method of claim 2 , wherein

the width of the blending area is based on N1 width candidates in the reordered width candidate list, N1 being less than a number of the width candidates in the width candidate list, TM costs corresponding to the N1 width candidates being less than or equal to one or more TM costs corresponding to one or more remaining width candidates in the reordered width candidate list.

8. The method of claim 2 , wherein:

the selected width candidate is a width candidate with the smallest TM cost in the reordered width candidate list.

9. A method for video decoding, comprising:

decoding coding information of a current block in a current picture, the coding information indicating that the current block is coded with a geometric partition mode (GPM) along a partition edge, a blending area that surrounds the partition edge being defined by boundaries on both sides of the partition edge, the boundaries being parallel to the partition edge, a width of the blending area measured perpendicular to the partition edge being based on a predefined width candidate list including width candidates;

determining a subset of width candidates based on the predefined width candidate list, a number of width candidates in the subset of width candidates being less than a number of the width candidates in the predefined width candidate list;

determining the width of the blending area by selecting a width candidate from the subset of width candidates; and

reconstructing samples within the blending area in the current block by applying a blending process using the selected width candidate.

10. The method of claim 9 , wherein the determining the subset of width candidates comprises:

selecting N2 smallest width candidates from the predefined width candidate list as the subset of width candidates when a screen content coding tool is enabled for a plurality of blocks including the current block.

11. The method of claim 9 , wherein the determining the subset of width candidates comprises:

determining the subset of width candidates based on a difference of two motion vectors of a first prediction mode and a second prediction mode used in the GPM to reconstruct the current block.

12. The method of claim 11 , wherein

screen content coding tools are disabled for the current block,

a first threshold is less than a second threshold, and

the determining the subset of width candidates includes:

determining the subset of width candidates as a first subset of width candidates when the difference of the two motion vectors is less than the first threshold;

determining the subset of width candidates as a second subset of width candidates different from the first subset when the difference of the two motion vectors is larger than or equal to the first threshold and is less than or equal to the second threshold; and

determining the subset of width candidates as a third subset of width candidates different from the first subset and the second subset when the difference of the two motion vectors is larger than the second threshold.

13. The method of claim 9 , wherein

the partition edge is extended into a current template with an extended portion in the current template, the current template including samples in a left neighboring reconstructed block or an above neighboring reconstructed block of the current block, and

the determining the subset of width candidates includes determining the subset of width candidates based on a difference of sample values across the extended portion of the partition edge, the samples being in the current template.

14. The method of claim 13 , wherein

a first sample difference threshold is less than a second sample difference threshold, and

the determining the subset of width candidates includes:

determining the subset of width candidates as a fourth subset of width candidates when an absolute value of the difference of sample values is less than the first sample difference threshold;

determining the subset of width candidates as a fifth subset of width candidates different from the fourth subset when the absolute value of the difference of sample values is larger than or equal to the first sample difference threshold and is less than or equal to the second sample difference threshold; and

determining the subset of width candidates as a sixth subset of width candidates different from the fourth subset and the fifth subset when the absolute value of the difference of sample values is larger than the second sample difference threshold.

15. The method of claim 9 , wherein

a partition of the current block is predicted by intra prediction, and

the determining the subset of width candidates includes selecting three largest width candidates from the predefined width candidate list as the subset of width candidates.

16. A method for video encoding, comprising:

when a current block is coded with a geometric partition mode (GPM) along a partition edge intersecting the current block and a condition for adaptive blending is satisfied,

reordering width candidates in a width candidate list using template matching (TM) that is based on a current template of the current block and reference templates that correspond to the respective width candidates;

selecting a width candidate from the reordered width candidate list;

determining a width of a blending area based on the selected width candidate, the blending area surrounding the partition edge as defined by boundaries on both sides of the partition edge, the boundaries being parallel to the partition edge, the width of the blending area measured perpendicular to the partition edge;

determining the blending area based on the width of the blending area; and

encoding samples within the blending area in the current block by applying the adaptive blending using the determined blending area.

17. The method of claim 16 , wherein the reordering the width candidates comprises:

extending the partition edge from the current block into the current template, an extended portion of the partition edge being in the current template;

for each width candidate in the width candidate list,

determining a reference template corresponding to the respective width candidate based on the GPM and the current template, reference samples within a template blending area that surround the extended portion of the partition edge being determined based on a blending process, the template blending area being in the reference template, a width of the template blending area being based on the respective width candidate; and

determining a TM cost corresponding to the respective width candidate based on the current template and the reference template; and

reordering the width candidate list based on the determined TM costs.

18. The method of claim 16 , further comprising

when the condition for the adaptive blending is not satisfied,

determining the width of the blending area based on the width candidate list without reordering the width candidates in the width candidate list; and

encoding the samples within the blending area in the current block by applying a blending process.

19. The method of claim 17 , wherein

the width of the blending area is based on N1 width candidates in the reordered width candidate list, N1 being less than a number of the width candidates in the width candidate list, TM costs corresponding to the N1 width candidates being less than or equal to one or more TM costs corresponding to one or more remaining width candidates in the reordered width candidate list.

20. The method of claim 17 , wherein:

the selected width candidate is a width candidate with the smallest TM cost in the reordered width candidate list.

Assignments (1)

Continuity (2)

References Cited (19)

Cited By (2)

US 12,513,306 US 12,621,464